Four new Ceratocystis spp. associated with wounds on Eucalyptus, Schizolobium and Terminalia trees in Ecuador

Transcription

1 Four new Ceratocystis spp. associated with wounds on Eucalyptus, Schizolobium and Terminalia trees in Ecuador Marelize van Wyk 1, Brenda D. Wingfield 1 and Michael J. Wingfield 1 (1) Department of Genetics, Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, Pretoria, 0002, South Africa Marelize van Wyk marelize.vanwyk@fabi.up.ac.za Abstract Species of Ceratocystis commonly infect wounds on trees. In this study, artificially induced wounds were made on the stems of Eucalyptus, Schizolobium and Terminalia trees in Ecuadorian lowland forests, in an effort to determine the presence of Ceratocystis spp. in that environment. Species belonging to the C. fimbriata sensu lato (s.l.) complex and others in the C. moniliformis s.l. complex were collected. Phylogenetic analyses for both major groups in Ceratocystis using three gene regions (ITS, β-tubulin and EF1-α) revealed three distinct clades in the C. fimbriata s.l. complex and two in the C. moniliformis s.l. complex. Isolates in the three clades representing the C. fimbriata s.l. complex represent morphologically distinct species that are described here as C. curvata sp. nov., C. ecuadoriana sp. nov. and C. diversiconidia sp. nov. Isolates in one of the two clades in the C. moniliformis species complex represented C. moniliformis sensu stricto (s.s.) and the other was of a species with a distinct morphology that is described here as C. sublaevis sp. nov. Introduction Ceratocystis spp. are perithcial Ascomycetes that commonly infect wounds or the cut surfaces of recently felled trees (Kile 1993; Barnes et al. 2003; Van Wyk et al. 2009). They are transferred to these infection sites by sap-feeding insects such as nitidulid beetles (Juzwik 1999). In addition, a small number of Ceratocystis species are associates of conifer-infesting bark beetles (Wingfield et al. 1997). Some species of Ceratocystis are important pathogens of fruit and forest trees or root crops (Kile 1993). In addition, many species are weakly pathogenic or saprophytes causing sap stain on timber. Most species that are weakly pathogenic or saprophytic reside in the C. moniliformis (Hedgc.) C. Moreau species complex (Hedgcock 1906; Van Wyk et al. 2006a) and some of the species related to C. coerulescens (Münch) B.K. Bakshi also cause sap stain particularly in conifers (Wingfield et al. 1997; Witthuhn et al. 1998). Species in the C. fimbriata Ellis & Halst. s.l. complex include important pathogens of trees causing vascular wilt and canker stain diseases (Roux et al. 1999; Engelbrecht et al. 2007; Roux and Wingfield 2009). Early studies treated Ceratocystis moniliformis and C. fimbriata as single species although it was recognised that C. fimbriata included substantial variability (Webster and Butler 1967). In recent years, numerous cryptic species that would previously have been accommodated in C. moniliformis have been described (Yuan and Mohammed 2002; Van Wyk et al. 2004, 2006a). Likewise, C. fimbriata s.l. accommodates a relatively large number of cryptic taxa (Van Wyk et al. 2004;

2 Engelbrecht and Harrington 2005; Johnson et al. 2005) with C. fimbriata s.s. defined by isolates phylogenetically related to those from sweet potato in the USA (Engelbrecht and Harrington 2005). It is likely that DNA sequence phylogenies and other emerging species concepts will reveal many other cryptic species in this group. In Ecuador, plantation forestry is based largely on Eucalyptus spp. (~50%) and Pinus spp. (~40%) but various other trees are being tested for their potential as plantation species. These include the non-native Terminalia ivorensis A. Chev. and the native Schizolobium parahybum (Vell.) S.F. Blake. One of the key considerations in evaluating species for plantation development in Ecuador concerns their health and in this regard, various disease problems have been encountered (Geldenhuis et al. 2004; Lombard et al. 2008). In a previous study, C. fimbriata s.l. and C. moniliformis were isolated from machete wounds on diseased S. parahybum in Ecuador (Geldenhuis et al. 2004). Given the importance of members of the C. fimbriata s.l. complex as tree pathogens (Kile 1993; Roux et al. 2004; Heath et al. 2009; Roux and Wingfield 2009), there has been concern that these fungi could present constraints to forestry in Ecuador. The tropical environment in areas of Ecuador where hardwood species are grown, results in a lush vegetation and considerable competition due to rapidly growing understory plants. These are typically cleared manually using machetes and physical wounds are common at the bases of trees. These wounds present ideal sites for infection by species of Ceratocystis s.l. Therefore, this study was conducted to determine which of these fungi might be present in the local environment. This was done by artificially wounding tree stems and collecting Ceratocystis spp. for identification. Materials and Methods Isolates Wounds were made on the stems of Eucalyptus deglupta Blume, Terminalia ivorensis and Schizolobium parahybum trees in Ecuador at three different sites and times. The first wounding trial was conducted on trees growing on the farms Rio Silanche and La Celica near Salinas in March 2004, where wounds were made on E. deglupta trees. A second trial was conducted at Rio Pitzara near Salinas during November 2005 and in this case, wounds were made on E. deglupta and T. ivorensis trees. A third trial at Rio Pitzara was conducted during February 2006 where wounds were induced on S. parahybum trees. Wounds were made on the stems of trees using a machete and were similar to those described by Barnes et al. (2003). Pieces of wood were collected from the treated trees approximately 4 weeks after wounding and fungi were isolated directly from structures on the wood surface or through carrot baiting (Moller and De Vay 1968). Spore droplets were removed from ascomatal necks, transferred to 2% (w/v) malt extract agar (MEA) (Biolab, Midrand, South Africa) supplemented with Streptomycin Sulphate (0.001 g vol 1, SIGMA, Steinheim, Germany) and incubated at 25 C until cultures sporulated. All isolates collected in this study are maintained in the culture collection (CMW) of the Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, South Africa and representative isolates have been deposited with the Centraalbureau voor Schimmelcultures (CBS), Utrecht, The Netherlands. Material consisting of dried cultures on 2% MEA of representative isolates were lodged with the National Fungal Herbarium (PREM), Pretoria, South Africa. DNA sequence comparisons DNA was extracted from selected isolates as described by Van Wyk et al. (2006b). PCR conditions and reactions for three gene regions; the Internal Transcribed Spacer (ITS) regions 1 and 2 including the 5.8S rrna operon, part of the Beta tubulin (βt) gene and part of the Transcription Elongation Factor 1-alpha (EF1-α), were as described by Van Wyk et al. (2006b). The primers

3 utilized in these PCR reactions were ITS1 & ITS4 (White et al. 1990) for the ITS region, βt1a & βt1b (Glass and Donaldson 1995) for the βt gene region and EF1F & EF1R (Jacobs et al. 2004) for the EF1-α gene region. Purification of these amplified products was achieved with 6% Sephadex G- 50 columns (Steinheim, Germany). The amplicons were sequenced in both directions using the ABI PRISMTM Big DYE Terminator Cycle Sequencing Ready Reaction Kit (Applied BioSystems, Foster City, California) as described in previous studies (Van Wyk et al. 2006b). Sequence and phylogenetic analyses The chromatographs were analysed using the freeware Cromas Lite 2.01 ( All isolates were subjected to ITS PCR and sequences were submitted to BLAST on NCBI ( for initial identification before they were selected for further sequencing. The sequences were subsequently divided into two groups based on the results of these analyses. They either grouped in the C. fimbriata s.l. complex or in the C. moniliformis s.l. complex. Sequence data for closely related species were taken from previous studies (Van Wyk et al. 2006a, 2007a, b, 2009) on these two groups of fungi. The sequences for the three gene regions for each of the two datasets (C. moniliformis s.l. and C. fimbriata s.l.) were aligned using MAFFT version ( (Katoh et al. 2002), and the alignments were confirmed manually in MEGA4 (Tamura et al. 2007). Phylogenetic analyses based on parsimony were done using PAUP version 4.0b10* (Swofford 2002). A Partition Homogeneity Test (PHT) was run to determine whether the sequences of three gene regions for each of the two datasets could be combined (Swofford 2002). Phylograms were generated using a heuristic tree search algorithm. Sequences were randomised (reps=1,000), Mulpar was effective and a TBR branch swapping algorithm was used in the search. Gaps were treated as a fifth character. Confidence intervals of branch nodes using 1,000 bootstrap replicates with the full heuristic search option were calculated. Ceratocystis virescens (R.W. Davidson) C. Moreau was used as the out-group taxon for both the datasets. All sequences derived from this study were deposited in GenBank (Table 1 and 2). MrModeltest2 (Nylander 2004) was used to determine the most appropriate model for the analyses of each gene region. The suggested partition-specific models were then included in the Bayesian analyses. Phylogenetic trees based on Bayesian inference were generated using MrBayes version (Ronquist and Huelsenbeck 2003). One million generations were run during which random trees were generated using the MCMC procedure. Four chains were applied and sampled every 100th generation. To avoid including trees that had been sampled before convergence, tree likelihood scores were assessed to determine the number of trees that were formed before stabilization. Trees outside the point of convergence were discarded by means of the burn-in procedure in MrBayes. The three gene regions of the five groups of isolates from Ecuador (three in the C. fimbriata s.l. group and two in the C. moniliformis s.l. group) were analyzed separately for each gene region (ITS, βt and EF1-α). The PAUP settings were the same as above for the combined dataset excluding an outgroup. To determine the amount of variation between the isolates from Ecuador, Molecular Evolutionary Genetics Analysis (MEGA) 4 (Tamura et al. 2007) was used. Fixed alleles were sought in each gene region of the respective groups. Allele networks were drawn using the software TCS (Clement et al. 2000) in order to observe the shared alleles between the three groups in the C. fimbriata s.l. dataset and the two groups in the C. moniliformis s.l. dataset. Table 1. Isolates and species within the C. fimbriata s.l. species complex used in this study.

4

5 Table 2. Isolates and species in the C. moniliformis s.l. species complex used in this study. Culture characteristics and morphology Culture morphology was assessed after 3 or 7 days of growth on 2% MEA (w/v) supplemented with Streptomycin Sulphate (100 mg per litre). Colony colour was assigned using the colour charts of Rayner (1970). Two distinctly different colony morphologies and colony colours were observed in the isolates from Ecuador when grown on 2% MEA. The two colony morphologies were representative of species in the C. fimbriata s.l. and C. moniliformis s.l. complexes. It was not possible to distinguish between isolates representing the C. fimbriata s.l. complex based on colony morphology. Three clades in this group were, however, identified based on phylogenetic inference. Three isolates representing each of these phylogenetic clades in the C. fimbriata s.l. group were selected for further morphological comparisons. The isolates representing species in the C. moniliformis s.l. complex had indistinguishable colony morphology. Two clades were identified in this group based on phylogenetic inference. One of these represented C. moniliformis s.s. and the other an undescribed taxon. Three isolates representing the latter group were thus selected for morphological characterisation. Optimum growth temperatures were determined by placing 4 mm plugs taken from the margins of actively growing cultures, mycelium side down, at the centres of 90 mm 2% MEA plates. Five replicates were made for each culture at seven different temperatures between 5 C and 35 C at 5 C

6 intervals. The growth was assessed 7 days after incubation for the isolates representative of species in the C. fimbriata s.l. complex. For isolates representing species in the C. moniliformis s.l. complex, growth was measured 3 days after incubation. For morphological characterisation, fungal structures, taken from 10 d-old cultures grown on 2% MEA were mounted in lactic acid. Photographic images were captured for all taxonomically important structures with a Carl Zeiss microscope and a Zeiss Axio Vision camera system. Three isolates were selected for each of the discrete phylogenetic clades in the C. fimbriata s.l. and C. moniliformis s.l. complexes. Fifty measurements were taken for all taxonomically useful characteristics for one isolate in each group and an additional 10 measurements were made for the remaining two isolates. Minimum, maximum, average and standard deviation (stdv) were calculated for each structure and these are presented as (minimum ) stdv minus the mean stdv plus the mean ( maximum). Results Isolates A total of 57 isolates representing Ceratocystis spp. were collected from the artificially induced wounds on trees. Of these, 42 were collected from E. deglupta, six were from T. ivorensis and nine were from wounds made on S. parahybum. Based on morphology, these isolates represented two discrete groups, one typical of species in the C. moniliformis s.l. complex and one in the C. fimbriata s.l. complex. Of these, nine in the C. moniliformis complex were from S. parahybum trees and two were from T. ivorensis trees. Forty two isolates in the C. fimbriata s.l. complex were from E. deglupta trees while four were from T. ivorensis trees. Phylogeny Two separate phylogenetic datasets were constructed. One of these was for isolates in the C. fimbriata s.l. complex and the other for isolates in the C. moniliformis s.l. complex. C. fimbriata s.l. species complex A P-value of 0.01 was obtained from the PHT for the dataset that included the three gene regions considered. Although this value was low, it indicated that the datasets could be combined (Sullivan 1996; Cunningham 1997). A total of 1,996 characters represented this dataset of which 1,087 were constant, 44 were parsimony uninformative and 865 characters were parsimony informative. Four most parsimonious trees were obtained after a heuristic search, one of which was selected for presentation (Fig. 1). The tree length and goodness of character fit indices were as follows: tree length=1,956 steps, consistency index (CI)=0.7, retention index (RI)=0.9 and a rescaled consistency index (RC) of 0.6. The HKY+G model was determined to be appropriate for the ITS gene region with the software MrModeltest2. The GTR+G model was best suited for the βt gene region, while the HKY+G model was found to be best suited for the EF1-α gene region. These model settings were included in the Bayesian analyses with 4,000 trees discarded because they were outside the point of convergence. The posterior probabilities of the branch nodes supported the bootstrap values of the selected tree obtained in PAUP (Fig. 1). Parsimony analysis placed isolates from Ecuador in three separate clades with 100, 96 and 100% support respectively (Fig. 1). The clades were designated as CF1, CF2, and CF3 on the phylogram and are, henceforth, referred to using this notation. All isolates in Groups CF1 and CF2 were from

7 E. deglupta with isolates CMW22432, CMW22433, CMW22435, CMW22442 residing in CF1 and isolates CMW22092, CMW22093, CMW22097, CMW22405 residing in Group CF2 (Fig. 1). Isolates in Group CF1 were phylogenetically distinct from all species in the C. fimbriata s.l. species complex with 100% bootstrap and Bayesian support. The isolates in Group CF1 were sister to C. colombiana M. van Wyk & MJ. Wingf., a pathogen of various tree crops in Colombia (Van Wyk et al. 2010a) (Fig. 1). Isolates residing in Group CF2 were phylogenetically distinct (Bootstrap 96%, Bayesian 98%) from all other species in the C. fimbriata s.l. species complex and formed a sister group to C. neglecta M. van Wyk, Jol. Roux & Rodas, a pathogen of Eucalyptus trees in Colombia (Rodas et al. 2008). Isolates in Group CF3 included those that were obtained from T. ivorensis trees (Fig. 1). This group was phylogenetically distinct from all species in this complex with 100% bootstrap and Bayesian support, with no sister group present. The tree topologies for each of the datasets of the seperate gene regions appeared similar when compared to the tree topology of the dataset for the combined gene regions (Fig. 2). The three seperate groups could still be identified, however, support for the branches was not as high as in the combined dataset. All three gene regions had fixed polymorphisms for all three groups within the C. fimbriata s.l. dataset. In the ITS region, there were 33 differences between CF and CF2, 52 differences between CF2 and CF3 and 64 differences between CF1 and CF3. In the BT region there were four differences between CF1 and CF2, nine differences between CF2 and CF3 and seven differences between CF1 and CF3. In the EF region, there were six differences between CF1 and CF2, 9 differences between CF2 and CF3 and nine differences between CF1 and CF3. The allele networks obtained showed that the three groups observed in the phylogenetic trees were also seen in the allele network as the single dataset of the three gene regions produced three separate allele networks (Fig. 3). C. moniliformis s.l. species complex The PHT for the isolates in the C. moniliformis complex gave a P-value of 0.01, and the datasets were thus combined (Sullivan 1996; Cunningham 1997). This dataset consisted of 1,824 characters of which 1,389 were constant, 11 parsimony-uninformative and 424 characters were parsimony informative. One of six trees obtained was selected for presentation (Fig. 4) and it had a length of 610 steps, a CI of 0.8, a RI of 0.9 and a RC of 0.8. With MrModeltest2, the HKY+I model was selected for the ITS dataset while the GTR+G model was selected for the βt gene region and the SYM+G model was best suited for the EF1-α gene region. These models were incorporated into the Bayesian analyses in MrBayes. Three thousand trees were discarded (burn-in) as they were outside the point of convergence. The posterior probabilities for the nodes were supportive of the bootstrap values obtained in PAUP (Fig. 4). Two well resolved clades emerged from the phylogenetic analyses. In one of these, referred to as CM1 (CMW22458, CMW22459, CMW22462, CMW22460) isolates from Ecuador grouped with C. moniliformis s.s. (61% bootstrap and 100% Bayesian) and they were all from S. parahybum (Fig. 4). Isolates (CMW22415, CMW22422, CMW22444, CMW22449) in the second clade (CM2) were phylogenetically distinct from all other isolates (97% bootstrap, 100% Bayesian) in the C. moniliformis s.l. species complex (Fig. 4). The latter isolates originated from both T. ivorensis and E. deglupta trees.

8

9 Fig. 1. Isolates of Ecuador obtained from E. deglupta and T. ivorensis trees in a phylogenetic tree based on the combined regions of the ITS, βt and EF1-α for species in the C. fimbriata s.l. species complex. Ceratocystis virescens represents the out-group taxon. Bootstrap values and Bayesian posterior probabilities (in brackets) are indicated at the branch nodes Analyses of the datasetes for the two groups in the C. moniliformis s.l. complex, based on single gene regions, gave results diferent to those emerging from the combination of the three gene regions (Fig. 5). For the ITS gene tree, there was no support for either of the two groups. For the βt gene region, all the isolates from Ecuador grouped in one clade. In the EF1-α gene tree, the two groups were well-defined with high bootsrap support and the tree had a topology similar to that for the combined gene sequences. Fixed polymorphisms were observed in two of the three gene regions considered. The βt gene region had one base pair difference while the EF1-α gene region had 14 differences between C. moniliformis and the isolates representing an apparently different species from Ecuador. There were no differences between isolates based on sequence comparisions for the ITS gene region (Van Wyk et al. 2006a, 2010b). The allele network obtained for the combined genes for the isolates in the C. moniliformis s.l. complex from Ecuador together with the C. moniliformis s.s. isolates is presented in Fig. 6. All isolates representing C. moniliformis s.s. grouped together in one cluster while the isolates representing the CM2 clade grouped seperately from them. Culture characteristics and morphology The isolates from Ecuador could be separated into two groups based on their culture morphology. The first group included isolates that grew slowly and had a greenish colour. The second group grew faster and were white to light brown with abundant aerial mycelium or little to no mycelium but abundant ascomata that covered the plates. The first group of isolates were morphologically most similar to species in the C. fimbriata s.l. complex and the remaining group was morphologically most similar to species in the C. moniliformis s.l. species complex. These two morphological groups were also consistent with the two groups emerging from DNA sequence comparisons. C. fimbriata s.l. species complex Based on phylogenetic analyses, three distinct groups (CF1, CF2, CF3) were identified in the C. fimbriata s.l. complex. However, only one group could be identified based on culture morphology. All isolates that represented C. fimbriata s.l. had greenish to brown colonies and a banana odour. They, however, differed from each other with regard to growth in culture. After 7 days of incubation, isolates representing Group CF1 (CMW22432, CMW22435 and CMW22442) had an optimum growth at 25 C (36 mm). No growth was observed at 5 C, 10 C or 35 C, limited growth (14 mm) occurred at 15 C and intermediate growth was observed at 20 C (28 mm) and 30 C (32 mm). Isolates representing Group CF2 (CMW22092, CMW22093 and CMW22097) had an optimum growth at 25 C (47 mm). No growth was observed at 5 C, 10 C and 35 C, limited growth was observed at 15 C (20 mm) and 30 C (24 mm) and intermediate growth was observed at 20 C (36 mm). Isolates representing Group CF3 (CMW22445, CMW22446 and CMW22447) had an optimum growth at 25 C (21 mm). No growth was observed at 5 C, 10 C or 35 C, limited growth was observed at 15 C (8 mm) and intermediate growth was observed at 20 C (16 mm) and 30 C (14 mm).

10

11 Fig. 2. Separate unrooted phylogenetic trees of the three groups in the C. fimbriata s.l. dataset representing the three gene regions. a ITS. b βt. c EF1-α. Bootstrap values are indicated on the branches. Fig. 3. Three separate allele networks obtained from one dataset, within the C. fimbriata s.l. complex, representing three gene regions (ITS, βt and EF1-α) of the isolates obtained from Ecuador on E. deglupta and Terminalia trees. The numbers represent CMW numbers as listed in Table 1. C. moniliformis s.l. species complex Based on phylogenetic analyses, two groups (CM1, CM2) of isolates were identified in the C. moniliformis s.l. species complex. However, no clear distinction could be made between them based on colony morphology. Both groups displayed very rapid growth, had white to light brown colonies with abundant aerial mycelium and a banana odour. Isolates representing Group CM1 (CMW22451, CMW22456, CMW22458, CMW22459, CMW22460, CMW22462) were identified as C. moniliformis s.s. based on the phylogenetic analysis and no further morphological characterisation was made for them. Isolates representing Group CM2 (CMW22444, CMW22449, CMW22410 and CMW22415) had an optimum growth after 3 days at 25 C (56 mm). No growth was observed at 5 C, 10 C or 35 C, limited growth was observed at 15 C (19 mm) and intermediate growth was observed at 20 C (41 mm) and 30 C (40 mm).

12

13 Fig. 4. Isolates of Ecuador obtained from S. parahybum, E. deglupta and T. ivorensis trees in a phylogenetic tree based on the combined regions of the ITS, βt and EF1-α for species in the C. moniliformis s.l. species complex. Ceratocystis virescens represents the out-group taxon. Bootstrap values and Bayesian posterior probabilities (in brackets) are indicated at the branch nodes Taxonomy The Ceratocystis isolates from Ecuador considered in this study clearly represent four previously unknown taxa, three in the C. fimbriata s.l. species complex and one in the C. moniliformis s.l. species complex. These differ from each other based on minor morphological characteristics and they are primarily distinguished based on DNA sequence comparisons for three gene regions. They are thus described as follows: C. curvata M. van Wyk & M.J. Wingf. sp. nov. (Fig. 7a i) MB Etymology: The name refers to the curved ascomatal necks in this species. Bases ascomatum nigrae globosae vel obpyriformes (162 ) ( 280) μm longae, (170 ) ( 280) μm latae. Colla ascomatum atrobrunnea vel nigra, apicibus undulates. Hyphae ostiolares divergentes, (34 ) ( 59) μm longae. Conidiophora biformia in mycelio singuli disposita, primaria phialidica hyalina basi tumescentia, apicem versus contracta (44 ) ( 139) μm longa; secondaria tubularia, expansa hyalina, (34 ) ( 90) μm longa. Chlamydosporae atro-brunneae, parietibus crassis, globosae vel subglobosae μm. Colony brownish olive (19 m) on malt extract agar. Mycelium submerged and aerial. Optimal temperature for growth 25 C, no growth below 15 C and above 35 C. Hyphae smooth, not constricted at septa. Ascomatal bases black, globose to obpyriform, no ornamentation, bases (162 ) ( 280) μm long, (170 ) ( 280) μm wide, (166 ) ( 280) μm in diam. Ascomatal necks dark brown to black, undulating at apices, (419 ) ( 714) μm long, (12 ) ( 25) μm wide at apices, (19 ) ( 56) μm wide at bases. Ostiolar hyphae divergent, hyaline, (34 ) ( 59) μm long. Asci not observed. Ascospores hat-shaped in side view, aseptate, hyaline, invested in a sheath, μm without sheath, μm with sheath. Ascospores accumulating in buff-yellow (19 d) mucilaginous masses at the apices of ascomatal necks. Thielaviopsis anamorph: Conidiophores of two types occurring singly on mycelium, primary conidiophores, phialidic, hyaline, swollen at the base, tapering towards the apices, (44 ) ( 139) μm long, 4 6 μm wide at bases, 4 7 μm wide at widest point in middle, 3 4 μm wide at apices. Secondary conidiophore, phialidic, flaring, hyaline, (34 ) ( 80) μm long, 3 6 μm wide at bases, 4 6 μm wide at apices. Conidia of two types: Primary conidia, hyaline, aseptate, cylindrical, (10 ) ( 31) 3 6 μm. Secondary conidia, hyaline, aseptate, barrel-shaped μm. Chlamydospores dark brown, thick walled, globose to perprolate, (7 ) 9 11 ( 12) μm. Habitat: Freshly wounded Eucalyptus deglupta wood. Known distribution: Ecuador. Material examined: Ecuador, near Salinas isolated from wood of E. deglupta, 2004, M.J. Wingfield (PREM60151 holotype), live strain: CMW22442, CBS

14

15 Fig. 5. Three separate unrooted phylogenetic trees of the two groups in the C. moniliformis s.l. dataset representing three gene regions. a ITS. b βt. c EF1-α. Bootstrap values are indicated on the branches. Where there are no bootstrap values this indicates that the bootstrap was less than 60%. Fig. 6. Allele network for representative isolates in the C. moniliformis s.l. complex. Isolates representing C. moniliformis s.s. and isolates obtained in this study from Ecuador were selected for this dataset. The ITS, βt and EF1-α gene regions were combined and a single allele network was obtained. The numbers in the network represent isolates in the CMW collection (Table 2)

16

17 Fig. 7. Morphological characteristics of Ceratocystis curvata. a Ascoma with globose to obpyriform base. b Ascomatal neck undulating at apex. c Divergent ostiolar hyphae. d Hat-shaped ascospores. e Primary conidiophore, flask-shaped phialides. f Secondary conidiophore, flask-shaped but flaring at apex. g Chain of cylindrical conidia. h Chain of barrel-shaped conidia. i Dark, globose to perprolate chlamydospores Bars; a = 100 μm. b, d, h = 5 μm. c, e, f, g, i = 10 μm Additional specimens examined (paratypes): same data (PREM60154), live strain CMW22432; same data (PREM60153), live strain CMW22433; same data (PREM60152), live strain CMW22435, CBS Notes: Ceratocystis curvata is distinct from all other species within the C. fimbriata s.l. complex in that it generally has ascomatal necks that are curved at the apices. Several other structures differ in size when compared to other species in the C. fimbriata s.l. complex (Table 3). C. ecuadoriana M. van Wyk & M.J. Wingf. sp. nov. (Fig. 8a h) MB Etymology: Name reflects the country where the fungus was first collected. Bases ascomatum nigrae globosae, (180 ) ( 290) μm longae, (184 ) ( 279) μm latae. Colla ascomatum atrobrunnea vel nigra, (515 ) ( 1,021) μm longa. Conidiophora biformia in mycelio singuli disposita, primaria phialidica hyalina basi tumescentia, apicem versus contracta (60 ) ( 98) μm longa; secondaria phialidica expansa hyalina, (52 ) ( 98) μm longa. Conidia biformia: primaria hyalina non septata cylindrica (9 ) ( 24) 3 5 μm. Colony brownish olive (19 m) on malt extract agar. Mycelium submerged and aerial. Optimal temperature for growth 25 C, no growth below 15 C and above 35 C. Hyphae smooth, not constricted at septa. Ascomatal bases black, globose, no ornamentation, bases (180 ) ( 290) μm long, (184 ) ( 279) μm wide, (182 ) ( 285) μm in diam. Ascomatal necks dark brown to black, (515 ) ( 1,021) μm long, (14 ) ( 28) μm wide at apices, (24 ) ( 46) μm wide at bases. Ostiolar hyphae divergent, hyaline, (29 ) ( 50) μm long. Asci not observed. Ascospores hat-shaped in side view, aseptate, hyaline, invested in sheath, μm without sheath, μm with sheath. Ascospores accumulating in buffyellow (19 d) mucilaginous masses on the apices of ascomatal necks. Thielaviopsis anamorph: Conidiophores of two types occurring singly on mycelium, primary conidiophore, phialidic, hyaline, swollen at the base, tapering towards the apices, (60 ) ( 98) μm long, 4 6 μm wide at bases, 5 7 μm wide at widest point in middle, 3 5 μm wide at apices. Secondary conidiophore, tubular, flaring, hyaline, (52 ) ( 98) μm long, 3 6 μm wide at bases, 4 6 μm wide at apices. Conidia of two types: Primary conidia, hyaline, aseptate, cylindrical, (9 ) ( 24) 3 5 μm. Secondary conidia, hyaline, aseptate, barrel-shaped 7 9 ( 11) 4 7 μm. Chlamydospores dark brown, thick walled, globose to sub globose, (9 ) ( 16)(7 ) 9 11 ( 12) μm. Habitat: Freshly wounded wood of Eucalyptus deglupta. Known distribution: Ecuador. Material examined: Ecuador, near Salinas isolated from wood of E. deglupta, 2004, M.J. Wingfield (PREM60155 holotype), live strain: CMW22092, CBS

18 Table 3. Morphological features of C. curvata, C. ecuadoriana and C. diversiconidia compared with each other and related species in the C. fimbriata sensu lato species complex. Additional specimens examined (paratypes): same data (PREM60156), live strain CMW22093, CBS124021; same data (PREM60158), live strain CMW22097, CBS124022; same data (PREM60157), live strain CMW Notes: There is no single morphological characteristic that distinguishes this species from other species within the C. fimbriata s.l. complex. It does, however, have very long ascomatal necks when compared to other species. Differences in the sizes and shapes of various other structures are also seen when compared to the other species within the C. fimbriata s.l. complex (Table 3).

19 Fig. 8. Morphological characteristics of Ceratocystis ecuadoriana. a Ascoma with globose base. b Divergent ostiolar hyphae. c Hat-shaped ascospores. d Primary conidiophore, flask-shaped phialides. e Secondary conidiophore, tubular flaring at apex. f Cylindrical conidia. g Chain of

20 barrel-shaped conidia. h Dark, globose to sub-globose chlamydospores Bars; a = 100 μm. c, f, h = 5 μm. b, d, e, g = 10 μm C. diversiconidia M. van Wyk & M.J. Wingf. sp. nov. (Fig. 9a i) MB Etymology: Name reflects the wide range of variation in size of both the cylindrical and barrel shaped conidia. Bases ascomatum nigrae globosae, (97 ) ( 196) μm longae, (97 ) ( 223) μm latae. Colla ascomatum atrobrunnea vel nigra, (245 ) ( 599) μm longa, basi (19 ) ( 34) μm lata. Hyphae ostiolares divergentes, hyalinae, (40 ) ( 72) μm longae. Conidiophora biformia in mycelio singuli disposita, primaria phialidica hyalina basi tumescentia, apicem versus contracta (58 ) ( 162) μm longa; secondaria phialidica expansa hyalina, (40 ) ( 68) μm longa. Colony brownish olive (19 m) on malt extract agar. Mycelium submerged and aerial. Optimal temperature for growth 25 C, no growth below 15 C and above 35 C. Hyphae smooth, not constricted at septa. Ascomatal bases black, globose no ornamentation, bases (97 ) ( 196) μm long, (97 ) ( 223) μm wide, (97 ) ( 210) μm in diam. Ascomatal necks dark brown to black, (245 ) ( 599) μm long, (12 ) ( 25) μm wide at apices, (19 ) ( 34) μm wide at bases. Ostiolar hyphae divergent, hyaline, (40 ) ( 72) μm long. Asci not observed. Ascospores hat-shaped in side view, aseptate, hyaline, invested in sheath, μm without sheath, μm with sheath. Ascospores accumulating in buff-yellow (19 d) mucilaginous masses on the apices of ascomatal necks. Thielaviopsis anamorph: Conidiophores of two types occurring singly on mycelium, primary conidiophores, phialidic, hyaline, swollen at the base, tapering towards the apices, (58 ) ( 162) μm long, 4 7 μm wide at bases, 4 9 μm wide at widest point in middle, 3 6 μm wide at apices. Secondary conidiophores, phialidic, flaring, hyaline, (40 ) ( 68) μm long, 4 6 μm wide at bases, 5 7 μm wide at apices. Conidia of two types: Primary conidia, hyaline, aseptate, cylindrical to oblong, apices truncate, (10 ) ( 30) 4 6 μm. Secondary conidia, hyaline, aseptate, barrel-shaped (5 ) 8 10 ( 11) 4 8 μm. Chlamydospores dark brown, thick walled, pyriform to obpyriform, (12 ) ( 19) (9 ) ( 15) μm. Habitat: Freshly wounded wood of Terminalia ivorensis. Known distribution: Ecuador. Material examined: Ecuador, near Salinas, isolated from wood of T. ivorensis, 2004, M.J. Wingfield (PREM60160 holotype), live strain: CMW22445, CBS Additional specimens examined (paratypes): same data (PREM60162), live strain CMW22448, CBS122605; same data (PREM60161), live strain CMW22446; same data (PREM60159), live strain CMW22447, CBS Notes: Ceratocystis diversiconidia is distinct from most species in the C. fimbriata s.l. complex due to its diverse conidial and chlamydospore sizes. There are also differences in sizes of various other structures when compared to other species in the C. fimbriata s.l. complex. (Table 3).

21

22 Fig. 9. Morphological characteristics of Ceratocystis diversiconidia. a Ascoma with globose base. b Divergent ostiolar hyphae. c Hat-shaped ascospore. d Primary conidiophore, flask-shaped phialides. e Secondary conidiophore, flask-shaped but flaring at apex. f Dark, pyriform to obpyriform chlamydospores. g Chain of barrel-shaped conidia with different sizes. h Primary conidia, cylindrical to oblong with truncated ends. Bars; a = 100 μm. b = 20 μm. c = 5 μm. d h = 10 μm C. sublaevis M. van Wyk & M.J. Wingf., sp. nov. (Fig. 10a h) Fig. 10. Morphological characteristics of Ceratocystis sublaevis. a Ascoma with globose base with limited amount of conical spines. b Divergent ostiolar hyphae. c Ascomatal neck disk-shaped at base. d One of a limited number of conical spines on ascomatal base. e Hat-shaped ascospores. f Primary conidiophore, flask-shaped phialides. g Primary conidia, cylindrical with truncated ends. h Secondary conidiophore, tubular flaring at apex. i Barrel-shaped conidia. Bars; a = 100 μm. b = 10 μm. c i = 5 μm MB Etymology: The name refers to the limited number of conical spines on the ascomatal bases that are somewhat smooth in comparison to other species in this group. Colonia in MEA alba vel laete brunnea. Mycelium abundans, plerumque aerium. Bases ascomatum nigrae globosae sparse spinis nigris et hyphis ornatae, spinae (2 ) 4 10 ( 13) μm, bases (98 ) ( 187) μm longae, (102 ) ( 231) μm latae. Colla ascomatum atrobrunnea vel nigra (100 ) ( 209) μm longa, discoidea vel papillata. Hyphae ostiolares divergentes, hyalinae,

23 (15 ) ( 25) μm longae. Conidiophora biformia in mycelio singuli disposita, primaria phialidica hyalina basi tumescentia, apicem versus contracta (15 ) ( 50) μm longa; secondaria tubularia apicem versus non contracta, hyalina rara (19 ) ( 36) μm longa. Conidia biformia: primaria hyalina non septata cylindrica μm; secondaria hyalina non septata doliiformia rara μm. Colony white to light brown (19 f) on malt extract agar. Mycelium abundant mostly aerial. Optimal temperature range for growth C, no growth below 15 C and above 35 C. Hyphae smooth, not constricted at septa. Ascomatal bases black, globose, ornamented sparsely with spines and hyphae, spines black, (2 ) 4 10 ( 13) μm long, bases (98 ) ( 187) μm long, (102 ) ( 231) μm wide, (100 ) ( 209) μm in diam. Ascomatal necks dark brown to black, undulating, (522 ) ( 990) μm long, (9 ) ( 19) μm wide at apices, (24 ) ( 55) μm wide at bases, with a disk to papillate base. Ostiolar hyphae divergent, hyaline, (15 ) ( 25) μm long. Asci not observed. Ascospores hat-shaped in side view, aseptate, hyaline, invested in sheath, μm without sheath, μm with sheath. Ascospores accumulating in buffyellow (19 d) mucilaginous masses on the apices of ascomatal necks. Thielaviopsis anamorph: Conidiophores of two types occurring singly on mycelium, primary conidiophores, phialidic, hyaline, swollen at the base, tapering towards the apices, (15 ) ( 50) μm long, 2 5 μm wide at bases, 2 5 μm wide at widest point in middle, 1 3 μm wide at apices. Secondary conidiophores, tubular, not-tapering towards apices, hyaline, scarce, (19 ) ( 36) μm long, 2 4 μm wide at bases, 2 4 μm wide at apices. Conidia of two types: Primary conidia, hyaline, aseptate, cylindrical, μm. Secondary conidia, hyaline, aseptate, scarce, barrelshaped μm. Habitat: Freshly wounded wood of Eucalyptus deglupta and Terminalia ivorensis. Known distribution: Ecuador. Material examined: Ecuador, Salinas area, isolated from freshly wounded wood of T. ivorensis, 2004, M.J. Wingfield (PREM60164 holotype), live strain: CMW22444, CBS Additional specimens examined (paratypes): Ecuador, isolated from wood of T. ivorensis, 2004, M.J. Wingfield (PREM60163), live strain CMW22449, CBS122517, Ecuador, isolated from wood of E. deglupta, 2004, M.J. Wingfield (PREM60165), live strain CMW22415; Ecuador, isolated from wood of E. deglupta, 2004, M.J. Wingfield (PREM60166), live strain CMW22422, CBS Notes: Ceratocystis sublaevis is morphologically distinct from all other species in the C. moniliformis s.l. complex. The most obvious distinguishing characteristic is found in its reatively small number of conical spines on the ascomatal bases. The very distinct disk-shaped structures at the bases of the ascomatal necks, typical of this group of fungi, is also very small in comparison to those seen in other species (Table 4). Discussion Five Ceratocystis spp. were isolated from freshly made wounds on the stems of plantation-grown trees in the lowland forests of Ecuador. Two of these species, C. sublaevis and C. moniliformis s.s. reside in the C. moniliformis s.l. species complex. The three remaining species reside in the C. fimbriata s.l. complex and they have been provided with the names C. curvata, C. ecuadoriana and C. diversiconidia. Both C. ecuadoriana and C. curvata strains were isolated from Eucalyptus trees, while C. diversiconidia was isolated from Terminalia trees. The C. moniliformis s.s. isolates were

24 isolated from Schizolobium while C. sublaevis was found on both Eucalyptus and Terminalia trees. Table 4. Comparison of morphological characteristics for C. moniliformis s.s., C. sublaevis and C. tribiliformis.

25 The three new Ecuadorian Ceratocystis spp. in the C. fimbriata s.l. complex reside in three distinct phylogenetic groups. Ceratocystis curvata in clade CF1 is most closely related to C. colombiana. The latter species is a pathogen of various tree crops (Van Wyk et al. 2010a) in Colombia that is geographically close to Ecuador where C. curvata was found. Ceratocystis ecuadoriana in clade CF2 was sister to C. neglecta that is a pathogen of Eucalyptus trees also in Colombia (Rodas et al. 2008). In contrast, C. diversiconidia in clade CF3 was distant from other species in the C. fimbriata s.l. complex without clearly defined phylogenetic neighbours. Based on phylogenetic inference, it should not be difficult to recognise these three new species from other taxa in the C. fimbriata s.l. complex that have been defined in previous studies (Johnson et al. 2005; Van Wyk et al. 2004, 2007b). It was not unusual to isolate C. moniliformis from wounds on trees in Ecuador as this fungus has previously been found in a similar habitat in the country (Geldenhuis et al. 2004). The new species, C. sublaevis, residing in the C. moniliformis s.l. complex was phylogenetically distinct from all species in this group. It was, however, closest related to C. moniliformis s.s. No differences were observed in the sequences of the ITS gene region when comparing isolates representing C. moniliformis s.s. and the isolates representing C. sublaevis. This was expected as it is known that, in contrast to isolates in the C. fimbriata s.l. complex, isolates in the C. moniliformis s.l. complex have no differentiation in the ITS gene region, while the EF1-α gene region has the highest degree of resolution (Van Wyk et al. 2006a, 2010b). The morphological characteristics of species in the C. fimbriata s.l. complex are very similar and new species can be recognised with confidence only through the application of phylogenetic inference. Yet, C. curvata has very distinct undulating ascomatal necks, C. diversiconidia has conidia and chlamydospores that differ significantly in size from other species in the group and C. ecuadoriana has unusually long ascomatal necks, which should assist in their identification. Ceratocystis sublaevis can be distinguished from other members of the C. moniliformis s.l. complex morphologically. The most distinct differences observed are the limited occurrences of conical spines on the ascomatal bases as well as its unusually small basal plateau giving rise to the ascomatal necks. Many Ceratocystis species are important plant pathogens (Kile 1993; Roux and Wingfield 2009). In this regard, nothing is known regarding the pathogenicity of the new Ceratocystis spp. described in this study. Given that there were no disease symptoms found on the wounded trees, it is possible that they all represent wound-infecting saprophytes. It would be worthwhile to consider their pathogenicity on the hosts from which they were isolated. In contrast, species in the C. moniliformis s.l. complex are typically non-pathogenic colonists of freshly wounded wood (Van Wyk et al. 2006a; Heath et al. 2009) and this is probably true for C. moniliformis s.s. and C. sublaevis found in this study. The discovery of a relatively large number of new taxa in a well-known group of fungi, reflects strongly the fact that these fungi have been poorly studied in Ecuador. It is likely that similar studies in that country will yield additional new species in this group and some of these are potentially important pathogens. Their discovery will enhance the understanding of fungal diversity in Ecuador and their description would augment a growing understanding of an intriguing and ecologically important group. Acknowledgments We are grateful to the National Research Foundation (NRF), members of the Tree Protection Cooperative Programme (TPCP), the THRIP initiative of the Department of Trade and Industry and the

26 Department of Science and Technology (DST)/NRF Centre of Excellence in Tree Health Biotechnology (CTHB) for funding. We also thank Mr Fernando Montenegro and Fundacion Forestal Manuel Durini for making collections of the fungi described in this study possible. Dr. Hugh Glen kindly suggested names for the new species and provided the Latin diagnoses. Literature Cited Barnes I, Roux J, Wingfield BD, Dudzinski MJ, Old KM, Wingfield MJ (2003) Ceratocystis pirilliformis, a new species from Eucalyptus nitens in Australia. Mycologia 95: Clement M, Posada D, Crandall K (2000) TCS: a computer program to estimate gene genealogies. Mol Ecol 9: Cunningham CW (1997) Can three incongruence tests predict when data should be combined? Mol Biol Evol 14: Engelbrecht CJB, Harrington TC (2005) Intersterility, morphology and taxonomy of Ceratocystis fimbriata on sweet potato, cacao and sycamore. Mycologia 97: Engelbrecht CJB, Harrington TC, Alfenas AC, Suarez C (2007) Genetic variation in populations of the cacao wilt pathogen, Ceratocystis cacaofunesta. Plant Pathol 56: Geldenhuis MM, Roux J, Wingfield MJ, Wingfield BD (2004) Development of polymorphic markers for the root pathogen Thielaviopsis basicola using ISSR-PCR. Mol Ecol Notes 4: Glass NL, Donaldson GC (1995) Development of primer sets designed for use with the PCR to amplify conserved genes from filamentous Ascomycetes. Appl Environ Microbiol 61: Heath RN, Wingfield BD, Wingfield MJ, Meke G, Mbaga A, Roux J (2009) Ceratocystis species on Acacia mearnsii and Eucalyptus spp. in eastern and southern Africa including six new species. Fungal Divers 34: Hedgcock GG (1906) Studies upon some chromogenic fungi which discolor wood. MO Bot Gard Annu Rep 17: Jacobs K, Bergdahl DR, Wingfield MJ, Halik S, Seifert KA, Bright DE, Wingfield BD (2004) Leptographium wingfieldii introduced into North America and found associated with exotic Tomicus piniperda and native bark beetles. Mycol Res 108: Johnson JA, Harrington TC, Engelbrecht CJB (2005) Phylogeny and taxonomy of the North American clade of the Ceratocystis fimbriata complex. Mycologia 97: Juzwik J (1999) Overland transmission of Ceratocystis fagacearum: extending our understanding. In: Cynthia L (ed) Shade tree wilt diseases: proceedings from wilt diseases of shade trees: a national conference. APS Press, St. Paul, pp Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res 30: Kile GA (1993) Plant diseases caused by species of Ceratocystis sensu stricto and Chalara. In: Wingfield MJ, Seifert KA, Webber JA (eds) Ceratocystis and Ophiostoma: taxonomy, ecology and

27 pathogenicity. APS Press, St. Paul, pp Lombard L, Bogale M, Montenegro F, Wingfield BD, Wingfield MJ (2008) A new bark canker disease of the tropical hardwood tree Cedrelinga cateniformis in Ecuador. Fungal Divers 31: Moller W, De Vay J (1968) Insect transmission of Ceratocystis fimbriata in deciduous fruit orchards. Phytopathology 58: Nylander JAA (2004) MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University. Rayner RW (1970) A mycological colour chart. Commonwealth Mycological Institute and British Mycological Society, Kew, Surrey. Rodas C, Roux J, Van Wyk M, Wingfield BD, Wingfield MJ (2008) Ceratocystis neglecta sp. nov., infecting Eucalyptus trees in Colombia. Fungal Divers 28: Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: Roux J, Wingfield MJ (2009) Ceratocystis species: emerging pathogens of non-native plantation Eucalyptus and Acacia species. Southern Forests 71: Roux J, Dunlop R, Wingfield MJ (1999) Susceptibility of elite Acacia mearnsii families to Ceratocystis wilt in South Africa. J For Res 4: Roux J, Van Wyk M, Hatting H, Wingfield MJ (2004) Ceratocystis species infecting stem wounds on Eucalyptus grandis in South Africa. Plant Pathol 53: Sullivan J (1996) Combining data with different distributions of among-site variation. Syst Biol 45: Swofford DL (2002) PAUP*. Phylogenetic analysis using parsimony (*and other methods). Version 4. Sunderland, Massachusetts: Sinauer Associates. Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol 24: Van Wyk M, Roux J, Barnes I, Wingfield BD, Liew ECY, Assa B, Summerell BA, Wingfield MJ (2004) Ceratocystis polychroma sp. nov., a new species from Syzygium aromaticum in Sulawesi. Stud Mycol 50: Van Wyk M, Roux J, Barnes I, Wingfield BD, Wingfield MJ (2006a) Molecular phylogeny of the Ceratocystis moniliformis complex and description of C. tribilliformis sp. nov. Fungal Divers 21: Van Wyk M, Van der Merwe NA, Roux J, Wingfield BD, Kamgan GN, Wingfield MJ (2006b) Population genetic analyses suggest that the Eucalyptus fungal pathogen Ceratocystis fimbriata has been introduced into South Africa. S Afr J Sci 102: Van Wyk M, Al-Adawi AO, Khan IA, Deadman ML, Al-Jahwari A, Wingfield BD, Ploetz R, Wingfield MJ (2007a) Ceratocystis manginecans sp. nov., causal agent of a destructive mango wilt